TypeScript transformer that generates run-time type-checks.
npm install --save typescript-is
# Ensure you have the required dependencies at compile time:
npm install --save-dev typescript
# If you want to use the decorators, ensure you have reflect-metadata in your depdendencies:
npm install --save reflect-metadata
If you've worked with TypeScript for a while, you know that sometimes you obtain any
or unknown
data that is not type-safe.
You'd then have to write your own function with type predicates that checks the foreign object, and makes sure it is the type that you need.
This library automates writing the type predicate function for you.
At compile time, it inspects the type you want to have checked, and generates a function that can check the type of a wild object at run-time. When the function is invoked, it checks in detail if the given wild object complies with your favorite type.
In particular, you may obtain wild, untyped object, in the following situations:
- You're doing a
fetch
call, which returns some JSON object. You don't know if the JSON object is of the shape you expect. - Your users are uploading a file, which is then read by your application and converted to an object. You don't know if this object is really the type you expect.
- You're reading a JSON string from
localStorage
that you've stored earlier. Perhaps in the meantime the string has been manipulated and is no longer giving you the object you expect. - Any other case where you lose compile time type information...
In these situations typescript-is
can come to your rescue.
NOTE this package aims to generate type predicates for any serializable JavaScript object. Please check What it won't do for details.
This package exposes a TypeScript transformer factory at typescript-is/lib/transformer-inline/transformer
As there currently is no way to configure the TypeScript compiler to use a transformer without using it programatically, the recommended way is to compile with ttypescript.
This is basically a wrapper around the TypeScript compiler that injects transformers configured in your tsconfig.json
.
(please vote here to support transformers out-of-the-box: microsoft/TypeScript#14419)
First install ttypescript
:
npm install --save-dev ttypescript
Then make sure your tsconfig.json
is configured to use the typescript-is
transformer:
{
"compilerOptions": {
"plugins": [
{ "transform": "typescript-is/lib/transform-inline/transformer" }
]
}
}
Now compile using ttypescript
:
npx ttsc
Please check the README of ttypescript for information on how to use it in combination with ts-node
, webpack
, and Rollup
.
There are some options to configure the transformer.
Property | Description |
---|---|
shortCircuit |
Boolean (default false ). If true , all type guards will return true , i.e. no validation takes place. Can be used for example in production deployments where doing a lot of validation can cost too much CPU. |
ignoreClasses |
Boolean (default: false ). If true , when the transformer encounters a class, it will ignore it and simply return true . If false , an error is generated at compile time. |
ignoreMethods |
Boolean (default: false ). If true , when the transformer encounters a method, it will ignore it and simply return true . If false , an error is generated at compile time. |
disallowSuperfluousObjectProperties |
Boolean (default: false ). If true , objects are checked for having superfluous properties and will cause the validation to fail if they do. If false , no check for superfluous properties is made. |
If you are using ttypescript
, you can include the options in your tsconfig.json
:
{
"compilerOptions": {
"plugins": [
{
"transform": "typescript-is/lib/transform-inline/transformer",
"shortCircuit": true,
"ignoreClasses": true,
"ignoreMethods": true,
"disallowSuperfluousObjectProperties": true
}
]
}
}
Before using, please make sure you've completed configuring the transformer.
In your TypeScript code, you can now import and use the type-check function is
(or createIs
), or the type assertion function assertType
(or createAssertType
).
For example, you can check if something is a string
or number
and use it as such, without the compiler complaining:
import { is } from 'typescript-is';
const wildString: any = 'a string, but nobody knows at compile time, because it is cast to `any`';
if (is<string>(wildString)) { // returns true
// wildString can be used as string!
} else {
// never gets to this branch
}
if (is<number>(wildString)) { // returns false
// never gets to this branch
} else {
// Now you know that wildString is not a number!
}
You can also check your own interfaces:
import { is } from 'typescript-is';
interface MyInterface {
someObject: string;
without: string;
}
const foreignObject: any = { someObject: 'obtained from the wild', without: 'type safety' };
if (is<MyInterface>(foreignObject)) { // returns true
const someObject = foreignObject.someObject; // type: string
const without = foreignObject.without; // type: string
}
Or use the assertType
function to directly use the object:
import { assertType } from 'typescript-is';
const object: any = 42;
assertType<number>(object).toFixed(2); // "42.00"
try {
const asString = assertType<string>(object); // throws error: object is not a string
asString.toUpperCasse(); // never gets here
} catch (error) {
// ...
}
You can also use the decorators to automate validation in class methods. To enable this functionality, you should make sure that experimental decorators are enabled for your TypeScript project.
{
"compilerOptions": {
"experimentalDecorators": true
}
}
You should also make sure the peer dependency reflect-metadata is installed.
npm install --save reflect-metadata
You can then use the decorators:
import { ValidateClass, AssertType } from 'typescript-is';
@ValidateClass()
class A {
method(@AssertType() value: number) {
// You can safely use value as a number
return value;
}
}
new A().method(42) === 42; // true
new A().method('42' as any); // will throw error
This family of functions check not only whether the passed object is assignable to the specified type, but also checks that the passed object does not contain any more than is necessary. In other words: the type is also "assignable" to the object. This functionality is equivalent to specifying disallowSuperfluousObjectProperties
in the options, the difference is that this will apply only to the specific function call. For example:
import { equals } from 'typescript-is';
interface X {
x: string;
}
equals<X>({}); // false, because `x` is missing
equals<X>({ x: 'value' }); // true
equals<X>({ x: 'value', y: 'another value' }); // false, because `y` is superfluous
To see the declarations of the functions and more examples, please check out index.d.ts.
For many more examples, please check out the files in the test/ folder. There you can find all the different types that are tested for.
- This library aims to be able to check any serializable data.
- This library will not check functions. Function signatures are impossible to check at run-time.
- This library will not check classes. Instead, you are encouraged to use the native
instanceof
operator. For example:
import { is } from 'typescript-is';
class MyClass {
// ...
}
const instance: any = new MyClass();
is<MyClass>(instance); // error -> classes are not supported.
// Instead, use instanceof:
if (instance instanceof MyClass) {
// ...
}
- This library will not magically check unbound type parameters. Instead, make sure all type parameters are bound to a well-defined type when invoking the
is
function. For example:
import { is } from 'typescript-is';
function magicalTypeChecker<T>(object: any): object is T {
return is<T>(object); // error -> type `T` is not bound.
}
If you stumble upon anything else that is not yet supported, please open an issue or submit a PR. 😉
Features that are planned:
- Promise support. Something like
assertOrReject<Type>(object)
will eitherresolve(object)
orreject(error)
. - Optimize the generated conditions. Things like
false || "key" === "key"
can be simplified. Might be more interesting to publish a different library that can transform a TypeScript AST, and then use it here, or use an existing one. Might be out of scope, as there are plenty of minifiers/uglifiers/manglers out there already.
git clone https://github.com/woutervh-/typescript-is.git
cd typescript-is/
npm install
# Building
npm run build
# Testing
npm run test